ZK Hints
Daily Digest每日摘要View All查看全部
Anthropic's team used Claude AI to reveal the origins of NIST ECDSA curve seeds, generated via Unix timestamps and the `lrand48` PRNG, dispelling long-standing backdoor concerns.Anthropic团队利用Claude AI揭示了NIST ECDSA曲线种子的来源，发现其由Unix时间戳和`lrand48`伪随机数生成器生成，消除了长期以来对潜在后门的疑虑。
Key Points:要点：
- •NIST ECDSA seeds generated via Unix timestamps and `lrand48`, not hashed phrasesNIST ECDSA曲线种子由Unix时间戳和`lrand48`生成，非哈希短语
- •Claude AI cracked the 25-year mystery by recognizing PRNG output patternsClaude AI通过识别PRNG输出模式破解了25年谜团
- •Simple seed generation process rules out backdoors种子生成过程简单透明，排除了后门可能性
- •Some base points generated via FreeLIP library's RNG部分基点由FreeLIP库的随机数生成器生成
- •B-163 seed and some base points remain unexplainedB-163种子及部分基点生成方式仍未解明
- •Findings bolster confidence in ECDSA curve security研究结果增强了对ECDSA曲线安全性的信心
@OpenAI released diverse mathematical results from an internal frontier model, hosted on GitHub with Lean formalizations, reasoning summaries, and compute statistics.@OpenAI 发布内部前沿模型生成的多项数学结果，托管于 GitHub 并提供 Lean 形式化证明，披露推理摘要与约三小时 ChatGPT Pro 算力统计。
Key Points:要点：
- •@OpenAI released many new mathematical results from an internal frontier model.@OpenAI 发布由内部前沿模型生成的多项数学新结果。
- •Results are hosted on GitHub with revision and citation protocols.结果托管在 GitHub，包含论文修订和引用协议。
- •Many proofs include Lean formalizations checkable by computer.许多证明提供 Lean 形式化，可由计算机检查。
- •It shares reasoning summaries, ChatGPT Pro compute estimates, and attempted-problem statistics.披露 10 份推理摘要、ChatGPT Pro 算力估算和问题尝试统计，平均约三小时 Pro 思考算力。
- •@OpenAI will fund workshops, conferences, and special programs, and continue evaluating frontier models.@OpenAI 将资助工作坊、会议和特别项目，并继续评估内部前沿模型。
Bozhko and Kutsenok (@reclaimprotocol) propose zkPoPE, a transport-agnostic zkTLS security model, and Reclaim Protocol, which UC-realizes it and attests 1200kB in 7.28s.Bozhko与 @reclaimprotocol 的Kutsenok在论文中提出应用层zkTLS安全模型zkPoPE，并给出UC实现Reclaim Protocol，可证明真实协议执行，支持1200kB响应证明。
Key Points:要点：
- •First transport-agnostic security model for zkTLS, defined at the application layer.首个传输无关的zkTLS安全模型，定义在应用层而非TLS层。
- •zkPoPE requires proof that private input and received reply satisfy a property, and that the reply came from a real execution, not fabrication.zkPoPE要求证明方证明其私有输入与真实收到的回复满足属性，且回复来自真实双方协议执行，而非伪造。
- •zkPoPE is formalized as UC ideal functionality F_zkpope.将zkPoPE形式化为UC理想功能F_zkpope。
- •Reclaim Protocol UC-realizes F_zkpope under a deployment-aligned trust model and underlies a production system.Reclaim Protocol在部署对齐信任模型下UC实现F_zkpope，并已支撑生产系统。
- •It introduces Distributed-AEAD, a new three-party authenticated-encryption primitive.引入Distributed-AEAD三方认证加密原语，作为协议构建基础。
- •On a lower-mid-range phone, it attests 1200kB in 7.28s, two orders beyond prior zkTLS benchmarks.中低端手机7.28秒证明1200kB响应，规模比此前zkTLS基准高两个数量级。
Eagen and Lai present Argo, a randomized encoding for invalidity of pairing-based SNARKs such as Groth16. An invalid proof lets a decoder reveal a secret and prove invalidity; combined with BitVM it enables a permissionless optimistic Groth16 verifier on Bitcoin. With BN254, encoding is about 21 kB.Eagen与Lai提出Argo，针对Groth16等配对SNARK的无效性随机编码。无效证明可让解码者公开秘密证明其无效，结合BitVM可在比特币上构建无需许可的乐观Groth16验证器，BN254编码约21 kB。
Key Points:要点：
- •Argo encodes invalid Groth16 proofs so a decoder can derive and reveal a secret.Argo将无效Groth16证明编码，使解码者可导出并公开秘密
- •Security relies on DDH in one source group of a bilinear pairing.安全性基于双线性配对单源群中的DDH假设
- •Invalidity Argo is selectively secure; validity Argo is adaptively secure.无效性Argo满足选择性安全，有效性Argo满足自适应安全
- •Argo MAC or Duty-Free Bits can build a Groth16 invalidity garbled circuit.可用Argo MAC或Duty-Free Bits构造Groth16无效性乱码电路
- •With BitVM, it enables a permissionless optimistic Groth16 verifier on Bitcoin.结合BitVM可实现比特币上无需许可的乐观Groth16验证器
- •For BN254, encoding is about 21 kB, over 1000x smaller end-to-end than Yao-based approaches.BN254下编码约21 kB，端到端规模比Yao方案小三个数量级
Akama et al. proposed zk-BAN, an anonymous blocklisting system with signature-based revocation. It filters revocation lists and partially unifies PRF tags to reduce repeated ZK computations. It proves blocklistability and unlinkability, achieving ~600 ms proving and ~2 ms verification under large-scale workloads.Akama等人提出匿名黑名单系统zk-BAN，通过过滤撤销列表和部分统一PRF标签减少重复ZK计算。系统支持基于签名的撤销，并证明满足可黑名单性与不可链接性。在大规模负载下，证明生成约600 ms、验证约2 ms，较先前工作证明快136倍、验证快32倍。
Key Points:要点：
- •Anonymous blocklisting revokes malicious users while preserving honest-user anonymity.匿名黑名单允许服务撤销恶意用户，同时保护诚实用户匿名性。
- •Signature-based revocation supports revocation from past authentications.基于签名的撤销支持根据过去认证记录撤销用户。
- •Prior systems incur high overhead at scale and for long-offline users.现有方案在大规模和长期离线用户场景开销较高。
- •zk-BAN filters revocation lists and partially unifies PRF tags to cut repeated ZK work.zk-BAN通过过滤撤销列表、部分统一PRF标签减少重复ZK计算。
- •The paper formally proves blocklistability and unlinkability.论文形式化证明zk-BAN满足可黑名单性与不可链接性。
- •It reports ~600 ms proving and ~2 ms verification, with 136x proving and 32x verification gains over prior work.大规模负载下证明约600 ms、验证约2 ms，较先前工作证明快136倍、验证快32倍。